This is a snapshot of some passion projects from many years ago, not long after the historic Colorado Flood in 2013.
I had posted a series of blog posts that focused on two professional interests: early warning systems and the Twitter developer platform. The posts started with a discussion of looking for rainfall 'signals' in Twitter data, then discussed potential roles the Twitter network could play in early-warning systems, and finally wrapped up with how the 2013 Colorado Flood unfolded on Twitter...
The data analysis behind these blog posts included a variety of software tools that helped harvest both Twitter and meteorological data, store that data in a MySQL database, and then perform the time-series analysis presented.
These tools fell out of this project:
- EventBinner: Ruby code that pulls data from MySQL and outputs event (Geo)JSON data. Combines Twitter and 'external' data into an integrated source file for project viewers.
- EventViewer: Web map-based event viewer. Source 'maperator' project lives here: https://github.com/blehman/maperator
- EventPlots: prototype for building d3 time-series plots of event.
- EventDeck: Receives streamed data and displays Tweets in pre-defined columns, ala TweetDeck.
